Comparison of Y-STR haplotypes in Colombian populations

To assess population structure across Colombian departments and regions, pairwise genetic distances (_F_ST and _R_ST) were calculated using data from this study and from previous publications that included at least the Y17 loci (Supplementary Tables S1 and S3). Departments with less than 16 haplotypes were excluded from these analyses: Casanare (n = 2), Meta (n = 1) and Caquetá (n = 6). Genetic distances were calculated for both 17 Y-STRs and 23 Y-STRs, with less populations available for the larger Y-STR set (Supplementary Table S1). The results obtained are presented in Supplementary Table S3 and represented in an MDS plot (Fig. 2).

In all analyses, Putumayo and Amazonas showed no significant differences among them, and appeared genetically distinct from all other departments, showing significant differences with all departments except Risaralda. This pattern is probably related to the higher Native American ancestry of the Amazonian populations when compared to other regions, which is supported by data from ancestry-informative markers [5].

A significant genetic differentiation from other departments was also evident for the department of Chocó, which only showed non-significant low differences when compared with Providencia and San Andrés for the 15 Y-STR set, probably due to shared African ancestry [8]. Studies using ancestry-informative markers also identified statistically significant differences between Andean populations (Antioquia, Bogotá, Norte de Santander) and Chocó [5, 65].

Among the Andean populations, significant differentiation was obtained for the department of Cauca, Nariño and Tolima in both _F_ST and _R_ST analyses for the smaller Y-STR set (Fig. 2A and C). It is worth noting that Nariño and Cauca showed the lowest genetic distances values with Putumayo, which is geographically close. The geographic proximity of these departments to the Amazon may have contributed to gene flow and, possibly, a greater Native American ancestry than the other departments in the Andean region. Indeed, a previous study using ancestry-informative markers showed that the southwestern Andean region, which comprises the departments of Nariño, Cauca, and part of Putumayo, preserved a significant proportion of Native American ancestry (37–44%) [5].

When comparing the five geographic regions of Colombia, the populations from Orinoquía, Andes, and Caribbean tend to cluster together (Fig. 2), with an overlap of the clusters of the Andes and Caribbean populations in the analyses performed for a smaller set of markers (Fig. 2A and C). In the analyses performed with 21 and 23 Y-STRs, a clearer separation of the five regions is observed (Fig. 2B and D).

Fig. 2: MDS plots based on FST and RST genetic distances. The number of loci varied in each analysis. In the FST analysis the complete sets of 17 Y-STRs (A) and 23 Y-STRs (B) were included, while in the RST analysis only 15 Y-STRs (C) and 21 Y-STRs (D) were used, by excluding DYS385 loci. Purple, gray and blue circles include population samples from the same Colombian region with non-statistically significant differentiation and genetic distances below 2%

Taken as a whole, the genetic distance results obtained demonstrate differences in paternal lineages between the five geographic regions of Colombia, which should be considered when evaluating forensic evidence. Although some differences were also observed between populations within the regions, the overall results did not reveal a statistically significant differentiation.

To assess the robustness of the groupings based on the genetic distances between pair of populations defined in Fig. 2, an AMOVA analysis was performed. Colombian departments included in the present study, together with previously published data [14, 16, 17, 29, 30], were grouped into six major regions: Insular, Andes, Caribbean, Orinoquía, Pacific, and Amazon for the analysis of 17-STRs and 23-STRs. The results obtained from the two sets of Y-STR markers indicated that majority of genetic variation occurred within populations (> 97.81%), while up to 1.53% of the variation was attributed to differences among groups, and less than 0.66% to differences among populations within groups.

Although the genetic differentiation between Colombian departments was low in both sets, it was statistically significant (_F_ST > 0.01914 and p = 0.00000). Also, values of among groups’ fixation index (_F_CT) were low (< 0.05), indicating small but statistically significant genetic differences among the regional groupings.

In the AMOVA based on 23 Y-STRs, values of among populations within groups’ fixation index (_F_SC) were not statistically significant (_F_SC = 0.00179, p = 0.16109). On the other hand, with the 17-STRs marker set, significant genetic differences were detected among populations within the regions (_F_SC = 0.00737, p = 0.00000). These findings suggest that an increase in the number of Y-STR markers appears to improve the resolution for identifying regional differences. Thus, based on the AMOVA results and genetic distance analyses, the regional groupings of the Colombian departments established for the 23 Y-STR markers set were adopted for further analyses.

Finally, it is worth noting that sampling in the Andean and Caribbean regions far exceeds that of Amazon, Pacific, and Orinoquía regions, reducing the statistical power and precision of estimates involving these regions. Therefore, future studies should increase sampling efforts in genetically unique but sparsely populated regions to achieve balanced sample sizes across regions, enabling more robust inter regional comparisons.